Baffled

In the past couple of years, scientists have spent an increasing amount of time being baffled. While there should be a certain degree of forgiveness for those baffled by the events of the past six months, it seems physicists in particular are baffled by results from experiments we designed and papers we have written. 

Maybe there is a minority of truly baffled scientists out there, progressing through their daily zoom meetings in a perpetual state of shock and awe. The popular media certainly makes it seem that we bounce around the pinball machine of the ivory tower from discovery to discovery without much preconception of what we are even looking for. It seems on an almost weekly basis a new headline alerts us to a new, not at all preempted or expected result. 

Science can be exciting. Science should be driven by the unexpected. But to imply that we are continuously in a state of bafflement about the results of experiments of our own design is disingenuous. It’s a part of modern science reporting that draws clicks and devalues public trust in what we do. 

Any inference (say a measurement) relies on our prior belief, and the method used to collect the data to make that inference also relies on prior belief. We don’t (in most cases) build experiments blindly. Often, an unexpected result is unexpected because it was simply lower down the list of probable events that could happen. 

Say you are told by weather forecasters that tomorrow there is an 80% chance that it will rain at some point during the day. You pack your umbrella and raincoat. Perhaps you drive to work. Your prior belief, informed by the weather forecast, is that it is more likely to rain than not. Consequently, when you head home and notice that the ground is completely dry and there is not a cloud in the sky, you are pleasantly surprised. Some of you may even be somewhat baffled. After all, the weather forecast told you that there was an 80% chance that there would be some rain at some point. 

In science, discovery is often driven by the 20% of the time it doesn’t rain. Usually, we didn’t make a better prediction of all the possible outcomes for two reasons: a lack of available data, or a model that inadequately represents reality. 

Yesterday’s announcement of an unusual compact binary coalescence (LIGO-speak for the event where a neutron star merges with either another neutron star or a black hole, or a black hole merges with another black hole) prompted multiple breathless reports of bafflement and amazement. However, just because the merger of two compact objects with vastly different masses (and one with a mass that challenges current ideas about compact objects) was unexpected, it does not mean that the measurement should not be trusted or should be chalked up to experimental errors. 

I am often the first person to critique analysis that draws expansive conclusions from tenuous ‘detections’, especially those heralded by publication in the highest impact journals. In this case, I would argue that any reports of bafflement are greatly overstated. It is widely known that astronomers understanding of how binary (and multiple) star systems evolve is a developing field of research. For the exact progenitor of GW190814 to be unclear at the moment is not without precedent: in the case of many Type Ia supernovae (the ones used for precision cosmology), we have no clear idea of what the progenitors of these events look like or how the explosion mechanism works. Closer to home, the progenitor of SN1987A was directly observed in archival images. The Blue supergiant progenitor was not predicted by stellar evolution models to result in SNe II at the time, however binary stellar evolution models reveal channels through which these stars can give rise to supernova explosions. 

It is worth pointing out that the observation of GW190814 does not rely on a single pipeline or analysis to derive the properties of the compact objects. The event was independently identified by all four online detection pipelines (including SPIIR, the pipeline I now work on. I was writing my research proposal for my current job when the event occurred!), all of which actually provide information on the chirp mass (amongst many other things) of the event. Further parameter estimation yields an exquisite analysis that many three-sigma discoveries in astrophysics should envy. 

It should go without saying that every event detected by LIGO undergoes rigorous vetting. It is nigh on impossible for a glitch from the detector to go as far as generating such a detailed analysis as has gone into this discovery.

While this observation is unexpected, it tells us something valuable, and that is that our current understanding of the physics of compact objects requires further investigation. I don’t think anyone involved in this field of research is truly baffled, which has become the new science reporting shorthand for a real eureka moment. And like a eureka moment, it’s far more likely our response is ‘that’s interesting’.